Home / Deep Research / Congenital Myasthenic Syndrome 15

Congenital Myasthenic Syndrome 15

Congenital Myasthenic Syndrome 15 Deep Research Fallback

⚠️ Fallback MONDO:0014542

Congenital Myasthenic Syndrome 15 Deep Research Fallback

Provider Attempts

  • openscientist: just research-disorder openscientist Congenital_Myasthenic_Syndrome_15 completed successfully in 1287 s and returned an 832-line, 9-citation report. The report was discarded without being curated from, because it is about the wrong disease.

Why the report was discarded

The report's central claim is:

Finding 1 — CMS15 is a presynaptic CMS caused by biallelic SLC18A3/VAChT variants

CMS15 (MONDO:0014542) is the ALG14 congenital myasthenic syndrome. MONDO defines the term as "Any congenital myasthenic syndrome in which the cause of the disease is a mutation in the ALG14 gene", with exact synonym ALG14 congenital myasthenic syndrome and cross-reference OMIM:616227. It was described alongside CMS14 (ALG2, with tubular aggregates) in a single 2013 study, PMID:23404334, and the MONDO synonym "without tubular aggregates" preserves that pairing. SLC18A3/VAChT congenital myasthenic syndrome is a separate entity; the 2025 review of CMS genes (PMID:40533459) lists ALG14 and SLC18A3 as distinct entries in the same gene list.

The report is internally coherent, well-cited science about a real disease. It is simply not this one, which makes it more dangerous than an obviously broken report — nothing in it looks wrong. It also asserts OMIM:616227, the correct OMIM number for CMS15, next to the wrong gene, so a curator spot-checking the identifier would find it matching.

just preflight-dr <report> MONDO:0014542 scored the report's gene mentions as SLC18A3=35, CHAT=10, DAP=4, CHRNE=3, RAPSN=3. ALG14 does not appear in the top five. The preflight nonetheless resolved to SKIP rather than a failure, because MONDO records no RO:0004003 causal gene for this term.

The report itself is deliberately not committed. research/ outputs are consumed as first-class curation inputs and indexed by disease name, so a wrong-disease report filed under the CMS15 name would mislead the next curator and the artifact index. This fallback record replaces it.

Two tooling gaps were filed:

  • dismech#9888 — just research-disorder passes a hardcoded empty mondo_id, so every deep-research run for every disease and provider is dispatched with no ontology grounding. That is the upstream cause of this misidentification.
  • dismech#9889 — preflight-dr silently skips its gene-identity check when MONDO has no RO:0004003 assertion, even where the MONDO definition and exact synonyms name the gene in plain text, as they do here.

Literature Scope Used Instead

Curation was anchored on generated PubMed reference caches:

  • PMID:23404334 — Cossins et al. 2013, Brain. The founding study identifying ALG14 and ALG2 as CMS genes; source for the limb-girdle phenotype, the ALG13/ALG14/DPAGT1 complex, endplate localisation of ALG14, and the siRNA experiment showing reduced surface acetylcholine receptor.
  • PMID:28733338 — Schorling et al. 2017, Neurology. Five patients from three families with the severe infantile form; source for the neurodegenerative pole of the spectrum, the recurrent p.Asp74Asn allele, recessive segregation, and the temporary-only pyridostigmine response.
  • PMID:34971077 — Long-surviving siblings with compound heterozygous ALG14 variants; source for endplate acetylcholine receptor deficiency, the decremental response, imaging findings, and the pyridostigmine benefit.
  • PMID:40533459 — 2025 review of the 40 genes causing CMS; used to confirm that ALG14 and SLC18A3 are separately recognised CMS genes.